Low-altitude airspace unmanned aerial vehicle group cooperative scheduling system and method based on multi-source data fusion

Abstract

The application provides a low-altitude airspace unmanned aerial vehicle group cooperative scheduling system and method based on multi-source data fusion, relates to the technical field of unmanned aerial vehicle group industrial control, and acquires fire field multi-source meteorological and thermal field distribution data, calculates a flow form adsorption index to calibrate a three-dimensional adsorption trap zone and a thermal convection dispersion zone. In the horizontal penetration stage, a guard unmanned aerial vehicle is scheduled to dynamically construct an asymmetric aerodynamic shielding volume on the windward side of the core machine, real-time monitoring of the fluid mechanics bearing limit of the single machine is performed, a near-death single machine centrifugal bowstring escape and a standby machine oblique cut-in replacement are triggered, and the array topology is maintained complete; in the vertical suppression stage, the vertical thermal convection speed and the load settlement critical value are coupled and solved, and the absolute negative pressure funnel effect inside the fire tornado is used to perform reverse suction throwing. The application overcomes the defect that the traditional single aircraft is easy to be rolled into a crash by strong shear wind, realizes safe penetration in an extreme thermal cyclone field, and improves the precision and efficiency of aerial fire extinguishing operation.

Description

Technical Field

[0001] This invention relates to the field of industrial control of unmanned aerial vehicle (UAV) swarms, specifically to a collaborative scheduling system and method for low-altitude UAV swarms based on multi-source data fusion. Background Technology

[0002] With the widespread adoption of drone technology, collaborative operations by heterogeneous drone swarms have become a crucial means of forest fire reconnaissance and suppression. To achieve precise delivery, the swarm must operate deep into high-risk low-altitude airspace. However, in massive forest fires occurring in complex terrains such as the southwestern mountainous regions, the combination of localized deflagration and valley winds can easily induce devastating extreme thermodynamic phenomena—thermal fire tornadoes. This extremely harsh physical environment presents unprecedented and catastrophic challenges to the existing control and coordinated dispatch of drone swarms.

[0003] Fire tornadoes are not simply strong winds, but three-dimensional fluid vortices with extremely strong tangential wind shear and centripetal negative pressure gradients. The extreme heat at the fire site causes an exponential decrease in local air density, resulting in a sharp drop in the aerodynamic lift of drone rotors. At this point, a deadly "manifold suction trap" forms around the fire tornado. When a swarm of drones approaches, traditional flight control logic such as "headwind deceleration and avoidance" or "hovering obstacle avoidance" instantly fails. Individual drones often cannot withstand the environmental suction due to the exhaustion of maximum thrust and are directly sucked into the vortex and crash. Under the impact of intense asymmetric turbulence, traditional conventional symmetrical drone formations are easily torn apart instantly. To escort the core drone carrying a heavy firefighting payload close to the fire source, independent flight control alone is no longer sufficient. In an extremely unstable thermodynamic boundary layer, how to utilize the hydrodynamic interference effect of multi-aircraft rotor wakes to dynamically create and maintain a safe "aerodynamic shielding" space for the core aircraft; and how to disrupt conventional flight control and utilize the centrifugal force of the wind field to achieve instantaneous escape and formation replacement when a single aircraft in the windbreak reaches its critical point of combat damage, are computational and mechanistic bottlenecks that current swarm collaborative algorithms struggle to overcome. Directly above a fire tornado exists an extreme thermal convection dispersion zone, where the raging vertical updraft speed often far exceeds the gravitational settling velocity of the fire extinguishing payload. If traditional gravity free-fall airdrops are used, the fire extinguishing bombs are easily deflected by strong winds or prematurely evaporate and fail under extremely high thermal radiation flux, leading to the complete failure of the core fire extinguishing mission. Summary of the Invention

[0004] The purpose of this invention is to provide a collaborative scheduling system and method for low-altitude airspace unmanned aerial vehicle (UAV) swarms based on multi-source data fusion, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion includes the following steps:

[0007] The system acquires a three-dimensional environmental topology map of the low-altitude airspace of forest fires, dynamic parameters of each aircraft in the UAV swarm performing firefighting and reconnaissance missions, a sequence of coordinated firefighting operations, and real-time environmental meteorological and combustion thermal field distribution data of the fire based on multi-source sensor fusion.

[0008] Based on real-time meteorological and thermal field distribution data of the fire scene, three-dimensional fluid dynamics and thermal dynamics spatial field modeling is used to mark the core area of ​​the three-dimensional vortex tube, the manifold adsorption trap area, and the extreme thermal convection divergence area of ​​the thermal fire tornado induced by the forest fire in the three-dimensional environmental topology map; the initial coordinated flight trajectory of the UAV swarm is obtained, and it is determined whether each node in the flight trajectory coincides with the manifold adsorption trap area; for conventional reconnaissance aircraft whose trajectories do not coincide, standard spatial obstacle avoidance path planning is executed to bypass the fire line; for core operational UAVs whose mission sequence requires them to cross or approach the manifold adsorption trap area to drop fire extinguishing payloads, the first asymmetric swarm penetration strategy is triggered, and multiple escort UAVs are dispatched. The human-machine interface dynamically constructs an aerodynamic shielding volume around the core operational drone. During the lateral penetration process in the first operational phase, the manifold adsorption index of each drone in the formation is calculated in real time. When the manifold adsorption index of the windward escort drone exceeds the preset limit bearing threshold, the second tangential slingshot escape and cluster formation self-healing strategy is triggered. The single drone executes the tangential slingshot escape trajectory to escape from the fire hazard zone, and the backup drone is simultaneously dispatched to replace the formation. In the vertical suppression operation in the second operational phase, when the core operational drone crosses the boundary of the manifold adsorption trap area and reaches the extreme thermal convection dispersion area, the escort cluster formation is disengaged, and a tangential negative pressure suction and throwing based on vertical airflow analysis is executed.

[0009] The low-altitude airspace UAV swarm collaborative scheduling system based on multi-source data fusion includes: a multi-source fusion and airspace topology reconstruction module, which is used to acquire a three-dimensional environmental topology map of the fire site and UAV dynamic parameters, fuse radar and multispectral data, perform spatial field modeling, and calibrate three major constraint boundaries: the three-dimensional vortex core region, the manifold adsorption trap region, and the extreme thermal convection divergence region.

[0010] The path cross-validation and intelligent diversion module is used to extract the initial trajectory of the UAV and perform overlap verification of trap zone nodes. It generates detour and obstacle avoidance corridors for routine tasks and triggers special operation strategy switching for core tasks.

[0011] The asymmetric aerodynamic shielding dynamic generation module is used to schedule the poses of multiple machines in the first operation phase, calculate the wind speed deficit based on the momentum conservation and wake thermal expansion mechanism, and generate and compensate the low wind speed safety envelope volume of the eccentric structure in real time.

[0012] The manifold adsorption monitoring and slingshot self-healing module is used to monitor the airframe's load-bearing limit at high frequency, hard block conventional flight control triggers tangential smooth acceleration, map the escape and jettison angle according to the topological curvature, and schedule redundant airframes to fill the gaps in the formation according to the penetration and damage rate.

[0013] The flow field game and negative pressure throwing control module is used to take over the flight control in the second operation phase, execute the cluster disbanding and wave-riding gliding commands, and control the payload to perform precise release and delayed detonation towards the center of the vortex negative pressure funnel by comparing the settling velocity with the dynamics of the updraft.

[0014] Compared with the prior art, the beneficial effects of the present invention are:

[0015] This invention calculates the manifold adsorption index by integrating multi-source fluid dynamics and thermodynamic parameters, and schedules a swarm of light UAVs to dynamically construct an asymmetric aerodynamic shielding volume on the windward side of the core UAV, so as to avoid individual UAVs being blown over by strong shear winds or engulfed in fire, thereby improving the penetration capability and overall survivability of heavy-load UAVs in extreme thermal cyclone fields.

[0016] This invention monitors the fluid carrying capacity limit of a single drone in real time. When encountering extreme turbulence, it uses the centrifugal force of a cyclone to throw away the endangered drone and simultaneously dispatches a backup drone to fill the aerodynamic gap. This avoids the energy depletion and crash caused by conventional hovering and avoidance, and improves the topology self-healing ability and collaborative operation fault tolerance of the drone swarm under complex thermal and wind fields.

[0017] This invention uses coupled calculation of vertical thermal convection velocity and load settling threshold to determine safe altitude, abandoning conventional gravity blind throwing. It utilizes the absolute negative pressure effect at the center of the fire tornado vortex tube to perform reverse suction throwing, avoiding the fire extinguishing agent being blown away by strong updrafts, and improving the targeting hit rate and effective agent coverage of aerial fire extinguishing loads in extreme convective environments. Attached Figure Description

[0018] Figure 1 The topological diagram of the low-altitude airspace UAV swarm based on multi-source data fusion constructing an asymmetric aerodynamic shielding formation in a manifold adsorption trap area and triggering the initial state of single-aircraft slingshot escape provided in the embodiments of the present invention.

[0019] Figure 2 The flow field geometry topology diagram provided by the embodiment of the present invention is for the strong shear wind field to penetrate and compress the aerodynamic shielding boundary, and for the method to plan the standby aircraft's leapfrog replacement trajectory;

[0020] Figure 3 The method provided in this embodiment of the invention completely shakes the windward side escort drone out of the danger zone, and the method is based on the instantaneous state diagram of the topology inclusion determination performed by the core operating machine in the absolute coordinate system;

[0021] Figure 4This invention provides a complete topology diagram illustrating the self-healing process of a drone swarm after the tangential slingshot escapes and is launched, with the backup drones obliquely cutting in to precisely fill the gaps in the formation.

[0022] Figure 5 This is a schematic diagram of the overall method and technical route of the present invention;

[0023] Figure 6 This is a schematic diagram of the overall system flow of the present invention. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Example 1:

[0027] Please see Figures 1 to 6 The present invention provides a technical solution:

[0028] The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion includes the following steps:

[0029] The system acquires a three-dimensional environmental topology map of the low-altitude airspace of forest fires, dynamic parameters of each aircraft in the UAV swarm performing firefighting and reconnaissance missions, a sequence of coordinated firefighting operations, and real-time environmental meteorological and combustion thermal field distribution data of the fire based on multi-source sensor fusion.

[0030] Based on real-time meteorological and thermal field distribution data of the fire scene, three-dimensional fluid dynamics and thermal dynamics spatial field modeling is used to mark the core area of ​​the three-dimensional vortex tube, the manifold adsorption trap area, and the extreme thermal convection divergence area of ​​the thermal fire tornado induced by the forest fire in the three-dimensional environmental topology map; the initial coordinated flight trajectory of the UAV swarm is obtained, and it is determined whether each node in the flight trajectory coincides with the manifold adsorption trap area; for conventional reconnaissance aircraft whose trajectories do not coincide, standard spatial obstacle avoidance path planning is executed to bypass the fire line; for core operational UAVs whose mission sequence requires them to cross or approach the manifold adsorption trap area to drop fire extinguishing payloads, the first asymmetric swarm penetration strategy is triggered, and multiple escort UAVs are dispatched. The human-machine interface dynamically constructs an aerodynamic shielding volume around the core operational drone. During the lateral penetration process in the first operational phase, the manifold adsorption index of each drone in the formation is calculated in real time. When the manifold adsorption index of the windward escort drone exceeds the preset limit bearing threshold, the second tangential slingshot escape and cluster formation self-healing strategy is triggered. The single drone executes the tangential slingshot escape trajectory to escape from the fire hazard zone, and the backup drone is simultaneously dispatched to replace the formation. In the vertical suppression operation in the second operational phase, when the core operational drone crosses the boundary of the manifold adsorption trap area and reaches the extreme thermal convection dispersion area, the escort cluster formation is disengaged, and a tangential negative pressure suction and throwing based on vertical airflow analysis is executed.

[0031] The low-altitude airspace UAV swarm collaborative scheduling system based on multi-source data fusion includes: a multi-source fusion and airspace topology reconstruction module, which is used to acquire a three-dimensional environmental topology map of the fire site and UAV dynamic parameters, fuse radar and multispectral data, perform spatial field modeling, and calibrate three major constraint boundaries: the three-dimensional vortex core region, the manifold adsorption trap region, and the extreme thermal convection divergence region.

[0032] The path cross-validation and intelligent diversion module is used to extract the initial trajectory of the UAV and perform overlap verification of trap zone nodes. It generates detour and obstacle avoidance corridors for routine tasks and triggers special operation strategy switching for core tasks.

[0033] The asymmetric aerodynamic shielding dynamic generation module is used to schedule the poses of multiple machines in the first operation phase, calculate the wind speed deficit based on the momentum conservation and wake thermal expansion mechanism, and generate and compensate the low wind speed safety envelope volume of the eccentric structure in real time.

[0034] The manifold adsorption monitoring and slingshot self-healing module is used to monitor the airframe's load-bearing limit at high frequency, hard block conventional flight control triggers tangential smooth acceleration, map the escape and jettison angle according to the topological curvature, and schedule redundant airframes to fill the gaps in the formation according to the penetration and damage rate.

[0035] The flow field game and negative pressure throwing control module is used to take over the flight control in the second operation phase, execute the cluster disbanding and wave-riding gliding commands, and control the payload to perform precise release and delayed detonation towards the center of the vortex negative pressure funnel by comparing the settling velocity with the dynamics of the updraft.

[0036] The application scenario in this example is a massive coniferous forest fire in the southwestern mountainous region, characterized by dramatic terrain and vegetation with high oil content. At 2:00 PM, influenced by valley winds and localized deflagration, a giant thermal fire tornado with a diameter of 150 meters was induced in the throat of the fire. Traditional single drones approaching the fire were sucked in and crashed. The command center urgently dispatched a drone swarm consisting of 30 highly mobile escort drones (Type A, 5kg weight) and 5 heavy firefighting core drones (Type B, 80kg payload) to carry out forced penetration and suppression of the core fire source.

[0037] S101. At time T0, the cluster ground control station acquires a WGS-84 3D environmental topology map of the target forest area via satellite imagery. Simultaneously, it reads cluster dynamic parameters, including:

[0038] Model A (Guardian Aircraft): Mass m A =5kg, maximum wind resistance thrust Fmax A =150N, maximum level flight speed Vmax A =35m / s.

[0039] Model B (Core Machine): Mass m B =120kg (full load), maximum wind resistance thrust Fmax B =450N, windward area S ref =1.5m 2 It simultaneously accesses real-time environmental meteorological and combustion heat field distribution data transmitted from meteorological radar and high-altitude reconnaissance aircraft.

[0040] S102. Divide the fire zone into a 2m×2m×2m three-dimensional voxel grid network. Substitute meteorological data into the Navier-Stokes fluid equation and the combustion thermodynamics equation.

[0041] Based on the calculation results, this method identifies three high-risk constraint zones on a 3D map (details of the calculation process are provided in subsequent embodiments): 3D vortex core zone: an absolute no-fly zone; Manifold adsorption trap zone: the area where the escort aircraft and core aircraft work together.

[0042] Extreme heat convection dispersion zone: directly above the fire, in the throwing operation area.

[0043] S103, Hierarchical Scheduling and Response Steps for Unmanned Aerial Vehicle (UAV) Swarms, specifically including:

[0044] For routine reconnaissance missions (a subset of aircraft type A), if the planned path is determined to pass through a manifold adsorption trap area, this method directly generates a safe corridor around the perimeter based on the A* optimization algorithm. For the bypass path planning of routine reconnaissance missions (aircraft type A) in this embodiment, this method does not employ the heuristic cost function optimization algorithm of the standard A* algorithm, which uses only three-dimensional Euclidean spatial distance as the single pathfinding cost. Instead, it deeply injects extreme thermodynamic and aerodynamic hyperparameters extracted from the aforementioned three-dimensional voxel grid network into the A* optimization algorithm, constructing a "multi-dimensional physical field coupling cost model" suitable for high-risk flow fields at the edge of fire tornadoes. Specifically, when calculating the movement cost g(n) of adjacent spatial grid nodes and the target approach prediction h(n), the system forcibly introduces four core environmental hyperparameters, including distance penalty, thermal convection disaster penalty, wind resistance dissipation penalty, and adsorption field gradient penalty, for dynamic weighted summation. Among them, the thermal convection-induced disaster penalty term calculates the approximation difference between the current grid's ambient temperature and the aircraft type A's ultimate tolerance temperature (preset 85℃) by reading real-time combustion heat field distribution data from weather radar. The closer the temperature is to the limit and the more severe the spatial temperature rise gradient of the grid, the more exponentially the pathfinding cost increases, thus forcibly avoiding the local deflagration high-temperature dispersion zone. The wind resistance dissipation penalty term directly calls the grid tangential wind speed three-dimensional vector calculated based on the Navier-Stokes fluid equation, performs a vector dot product operation with the planned flight orientation of the aircraft type A, thereby accurately quantifying the headwind drag and crosswind shear on the maximum wind-resistant thrust F of the aircraft type A. max The kinetic energy consumption ratio of (150N) is such that when the local turbulence variance or wind resistance exceeds the power margin threshold, the grid node will be given an extremely high resistance coefficient; the adsorption field gradient penalty term extracts the MAI (ambient suction to maximum thrust ratio) distribution tensor of the manifold adsorption trap area, and constructs a spatial repulsive force field similar to an artificial potential field, so that the grid pathfinding cost increases sharply as its distance from the adsorption trap area boundary decreases. Through the above dynamic hyperparameter injection based on real-time data of the fire physical environment, the A pathfinding cursor in this method no longer blindly searches for the geometric shortest path when facing a complex valley with wind and fire, but can adaptively perceive and avoid the dangerous boundaries of the high-temperature heat plume core, the severe wind shear section and the adsorption trap. Under the premise of strictly ensuring the power limit and thermal protection safety boundary of the aircraft A, it obtains a three-dimensional bypass reconnaissance corridor that is close to the periphery of the high-risk area but absolutely safe along the valley with the lowest comprehensive physical cost.

[0045] For the core firefighting mission (1 Type B aircraft + 6 Type A aircraft), the mission command requires that the fire extinguishing bombs be delivered to the ignition point at the bottom of the fire tornado's center. This method determines that the planned flight path must pass through the manifold adsorption trap zone. Immediately interrupt the obstacle avoidance logic of the conventional PID flight control, triggering the "first asymmetric swarm penetration strategy" (see Example 3); monitor the limits of individual aircraft in real time during flight, triggering "single-aircraft slingshot escape" (see Example 4) and "dynamic formation replacement" (see Example 6).

[0046] Example 2:

[0047] A fire tornado is not simply a strong ambient wind, but an extreme thermal cyclone field with strong centripetal suction and tangential wind shear. If the radial suction force exerted by this wind field on an aircraft is not precisely quantified, a drone will be sucked in and crash due to power exhaustion as it approaches its outer edge. This embodiment demonstrates how to sequentially calculate environmental parameters based on principles of fluid dynamics and thermodynamics, derive the manifold adsorption index (MAI), and delineate manifold adsorption trap regions.

[0048] The process of calibrating the manifold adsorption trap zone includes: extracting real-time environmental meteorological and combustion thermal field distribution data of the fire throat or concentrated combustion zone of the forest fire at the current timestamp; identifying the center coordinates of the rotating wind field of the thermal fire tornado induced by the convergence of strong ground thermal convection and environmental wind field through three-dimensional fluid dynamics modeling; and calculating the vortex core radius, centripetal negative pressure gradient, and tangential wind speed vector of each spatial grid. Three-dimensional spatial grids with a distance from the center coordinates smaller than the vortex core radius and a central airflow exhibiting a strong spiral upward state are marked as three-dimensional vortex core regions and considered as such. Absolute no-fly zone; For the spatial grid surrounding the core area of ​​the three-dimensional vortex tube, the manifold adsorption index (MAI) is calculated based on the centripetal negative pressure gradient, the abrupt change in air density caused by high temperature, and the upper limit of the maximum wind resistance thrust in the dynamic parameters of the UAV swarm. The physical characterization of the calculation formula is: the ratio of the centripetal negative pressure gradient to the product of the fire air density and the square of the upper limit of the maximum wind resistance thrust; The set of continuous spatial grids with a manifold adsorption index (MAI) greater than the preset critical adsorption threshold (MAI>1.0) and a continuously increasing tangential wind speed vector is marked as the manifold adsorption trap zone.

[0049] S201. Using a multi-source data fusion method combining multispectral images of the fire site with Doppler weather radar, the highest temperature gradient convergence point is identified as the coordinate of the wind field center (X). c ,Y c Z c ).

[0050] Based on the principle of conservation of angular momentum in fluid mechanics, the tangential wind speed of each spatial grid outside the vortex core region is calculated. :

[0051]

[0052] in, Γ is the tangential wind speed at a radius r from the center (unit: m / s); Γ is the calculated environmental wind field circulation constant (unit: m). 2 / s); r is the radial distance (in meters) from the spatial grid to the vortex center; the closer to the cyclone center, the more rapidly the angular velocity and linear velocity of the fluid rotation increase inversely, forming a high-speed tangential shear wind field on the periphery. The measured radius R of the fire tornado vortex core was... core =25m, maximum tangential wind speed V at the boundary θmax =32m / s.

[0053] S202. Due to the enormous heat released by the fire, the ambient air density drops sharply, which will directly cause a significant decrease in the drone's rotor lift. The local air density of each spatial grid is calculated based on the thermodynamic gas law:

[0054]

[0055] Where, ρ fire Local air density at the fire site (unit: kg / m³) 3 P0 is the ambient baseline atmospheric pressure, taken as 101325 Pa or kg / (m·s). 2 ); R spec T is the gas constant, with a value of 287 J / (kg·K); fire This represents the measured absolute temperature (in K) of the corresponding spatial grid. In an open fire zone, although the local atmospheric pressure is roughly balanced, extreme high temperatures cause the gas to expand violently, resulting in a sharp decrease in density. For example, when the measured temperature T of the grid... fire At 800K (approximately 527℃), ρ is calculated. fire ≈0.44kg / m 3 Only in a normal temperature environment (1.22kg / m³) 3 One-third of ).

[0056] Based on the principle of fluid centripetal force balance, the centripetal negative pressure gradient is calculated:

[0057]

[0058] in, Centripetal negative pressure gradient (unit: Pa / m or N / m) 3 The technical principle is that the rotating fluid itself will generate an outward centrifugal tendency. In order to maintain the rotation of the flow field and prevent it from collapsing, a continuous pressure difference (pressure gradient) is formed from the external high-pressure zone to the internal low-pressure zone. This pressure difference acts on the drone's body, macroscopically forming a negative pressure suction force that strongly pulls the drone into the core area of ​​the vortex.

[0059] S203. Extract the maximum wind-resistant thrust limit F of the escort drone (model A) performing the mission. max =150N. Based on the aerodynamic drag equation, it is equivalently transformed into the square term of the maximum wind resistance speed under low-density fire conditions. :

[0060]

[0061] Among them, S ref Reference area of ​​the aircraft facing the wind (unit: m²) 2 ); C d This is the air drag coefficient (dimensionless).

[0062] Based on the aforementioned relationship between environmental hazards and airframe resistance, the formula for calculating the manifold adsorption index (MAI) is derived as follows:

[0063]

[0064] The dimension of the numerator ∇Psuction(r)·r is (N / m3)·m=N / m 2 =Pa (pressure); the denominator represents the maximum propulsive dynamic pressure of the UAV, with dimensions (kg / m²). 3 )·(m 2 / s 2 )=N / m 2 =Pa (pressure). Dividing the two, MAI is a dimensionless number; the technical principle of the formula is that the numerator represents the equivalent value of the work done by the radial suction force imposed on the fuselage by the external flow field; the denominator represents the ultimate aerodynamic pressure resistance that the UAV can output after exhausting all its electrical energy. This index quantifies the ratio between "environmental traps" and "aircraft limits" from the bottom layer of fluid physics;

[0065] S204. The MAI value calculated for each grid in the three-dimensional space is compared with the preset critical adsorption threshold. The preferred range of the critical adsorption threshold is set to 0.85 to 1.15 (preferred value is 1.0 or a dynamic safety redundancy value set based on the UAV model). The preferred range given in this invention aims to balance early warning sensitivity and false alarm rate.

[0066] When MAI > critical adsorption threshold, it indicates that even with the UAV escaping at full power, it cannot counteract the environmental suction (even with maximum reverse thrust, it still cannot overcome the centripetal adsorption force). A continuous set of spatial grids where MAI > critical adsorption threshold and the tangential wind speed vector continuously increases is strictly marked as a manifold adsorption trap region in the 3D topology map. The corresponding data calculations are shown in Table 1.

[0067] Table 1: Calculation of Thermal and Manifold Adsorption Index (MAI) at Different Radius from Fire Center (for Model A)

[0068]

[0069] As shown in Table 1, at a distance of 28 meters from the cyclone's center, the MAI index undergoes a sharp change and exceeds the 1.0 boundary due to the sharp decrease in air density and the surge in wind speed caused by extremely high temperatures. If traditional wind speed measurement obstacle avoidance logic is used, the aircraft might remain unaware at 40 meters, only to lose control at 28 meters due to loss of lift and suction overload. This method predefines the physical trap boundary, providing a core basis for subsequently triggering swarm penetration strategies.

[0070] The technical principle of this embodiment lies in abandoning the superficial logic of traditional UAVs relying solely on wind speed thresholds for obstacle avoidance, and instead starting from the underlying physical mechanisms of three-dimensional fluid dynamics and combustion thermodynamics for swarm scheduling. In the extreme environment of a massive forest fire triggering a thermal fire tornado, the high temperature of the fire causes a drastic decrease in local air density, resulting in a sharp drop in the aerodynamic lift of the UAV rotor; simultaneously, the high-speed rotating vortex flow field generates an inward negative pressure gradient, forming a strong centripetal suction. This method first integrates multi-source sensor data to construct a three-dimensional voxel grid map, extracts the real-time temperature of each grid to calculate the attenuated local air density, and combines the principle of conservation of angular momentum to calculate the tangential wind speed and the centripetal negative pressure gradient, thereby obtaining the equivalent value of the centripetal work exerted on the fuselage by the environmental flow field. This method calculates the ratio of this environmental suction value to the limiting aerodynamic thrust dynamic pressure that the UAV can output under the current low-density air, deriving a dimensionless manifold adsorption index. This index precisely quantifies the relationship between the aircraft's wind resistance limit and environmental traps at the physical level. When this index exceeds a preset critical value, it means that even if the UAV escapes at full power, it cannot counteract the centripetal environmental suction. Based on this core index, this method pre-defines the manifold adsorption trap zone and the absolutely no-fly zone of the vortex core in the topology map. During the penetration operation phase, for the core UAV that must pass through the trap zone to drop fire extinguishing payloads, this method hard-blocks the conventional flight control's deceleration and hovering obstacle avoidance logic, because hovering in the suction trap is equivalent to direct crash. This method dispatches multiple light escort UAVs in a dense formation on the windward side of the core UAV, using rotor wake interference to dynamically construct an asymmetric aerodynamic shielding envelope to protect the core UAV from being torn apart by strong shear winds. During the journey, this method calculates the manifold adsorption index of the edge escort UAVs at high frequency. Once it approaches the load limit, it immediately uses the centrifugal force field generated by the wind rotation to guide the endangered single UAV to escape tangentially, and simultaneously dispatches a backup UAV in a safe redundancy state from the leeward side to fill the gap, maintaining the integrity of the shielding volume, until the core UAV reaches directly above the vortex eye to perform a tangential negative pressure suction drop. The beneficial effects of this embodiment lie in solving the problem of large-scale loss of control and crashes caused by sudden drops in lift and negative pressure adsorption when traditional UAV swarms encounter complex and variable flow fields such as thermal fire tornadoes. By innovatively constructing a manifold adsorption index, this scheduling method can accurately predict the survival critical point of individual UAVs at a multi-dimensional physical level, upgrading traditional spatial geometric obstacle avoidance to a deep thermal and hydrodynamic energy game. This method of pre-defining physical limit boundaries avoids the sluggishness and failure of conventional wind speed measurement obstacle avoidance logic in extremely high temperature and low pressure environments. The hierarchical collaborative scheduling strategy for heterogeneous swarms enables lightweight, highly maneuverable UAVs to act as dynamic protective barriers for heavy core operational UAVs. Through the perfect combination of asymmetric windproof formations and individual UAV slingshot escape and dynamic replacement mechanisms, the overall topology of the swarm is ensured not to be broken through by extreme turbulence.This enables drone swarms to penetrate extremely dangerous areas that are completely inaccessible to traditional aircraft, ensuring that critical firefighting payloads can be safely escorted to the throat of the fire and accurately delivered, thereby improving the efficiency of aerial firefighting and the overall survival rate of equipment in complex and harsh forest fire environments.

[0071] Example 3

[0072] The determination of extreme thermal convection divergence zones includes: extracting the vertical heat flux and vertical updraft velocity directly above the center of the forest crown fire based on real-time environmental meteorological and combustion heat field distribution data; calculating the predicted terminal free settling velocity of the mission payload at each altitude layer by combining the aerodynamic parameters of the mission payload, such as fire extinguishing bombs or liquid fire extinguishing agents carried by the core operation UAV; and marking the three-dimensional spatial range where the vertical updraft velocity is greater than the predicted terminal free settling velocity of the mission payload, and the evaporation rate of the fire extinguishing payload due to high-temperature baking exceeds a preset threshold, as the extreme thermal convection divergence zone. Specifically, this includes:

[0073] S205. Obtain the real-time absolute temperature T at different heights z directly above the center of the forest crown fire. fire (z) Due to the concentrated deflagration on the ground, the local air is extremely heated, resulting in a sharp decrease in density and the formation of a powerful buoyancy-driven flow in the gravitational field. This method is based on the thermodynamic plume core region equation to calculate the vertical updraft velocity V within the three-dimensional vortex eye directly above the fire. updraft (z):

[0074]

[0075] Where z is the vertical height from the center of the fire source in the tree canopy (unit: m), and g is the acceleration due to gravity (9.8 m / s²). 2 ); The measured temperature of the grid is located at a vertical distance from the center of the fire source in the tree canopy. The safe baseline atmospheric temperature outside the fire area is 25°C, which corresponds to 298K. The value of 0.85 is the plume turbulence correction factor, which is taken as 0.85 at the center of the fire tornado. The physical meaning is that the greater the temperature difference, the stronger the chimney effect, and the upward dynamic pressure wind speed rises nonlinearly and drastically.

[0076] S206. Introduce aerodynamic parameters of the fire extinguishing load and calculate the terminal free settling velocity: Extract the physical and aerodynamic parameters of the 80kg dry powder fire extinguishing bomb (or liquid fire extinguishing pack) carried by the core operation UAV, including: load mass m payload Load frontal cross-sectional area A ref air drag coefficient C of the load dAnd considering the gravitational acceleration constant g, when the load detaches from the drone and falls towards the throat of the fire tornado, it is theoretically subject to a struggle between a constant downward gravitational force and upward fluid resistance. Based on Newton's second law and the aerodynamic drag equation, the load's velocity ρ at various altitude levels is calculated. fire (z) Predicted terminal free settling velocity V at air density terminal (z) (i.e., the maximum falling speed when gravity and drag are in equilibrium):

[0077]

[0078] The principle behind the formula is that, as the temperature rises and the air density decreases towards the bottom of the fire center, the fire extinguishing bomb should theoretically fall faster; however, the violent updrafts coming from below increase exponentially. This is the key factor determining whether the load can penetrate the fire.

[0079] S207. Calculation of Load Evaporation / Loss Rate Based on Stefan-Boltzmann Law: Besides physical deflection, extreme high-temperature thermal radiation can cause liquid extinguishing agents to vaporize or dry powder to solidify and fail. This method extracts the thermal radiation flux at each height level directly above the fire. Based on Stefan-Boltzmann radiation law, the evaporation rate E of the extinguishing load per unit time as it falls through the spatial grid is calculated. rate (z):

[0080]

[0081] in, Let be the equivalent blackbody radiation emissivity of the fire scene, and σ be the Stefan-Boltzmann constant, 5.67 × 10⁻⁶. 8 W / (m 2 ·K 4 );S surface H is the surface area of ​​the load; v This refers to the latent heat of vaporization or heat of thermal decomposition of the extinguishing agent (unit: J / kg). This formula calculates the mass (kg / s) lost per second by the load due to "thermal vaporization" at high temperatures.

[0082] S208. Perform a joint Boolean determination on the environmental parameters and load limits calculated for each spatial grid. A three-dimensional spatial grid is considered valid if it satisfies any of the following conditions:

[0083] Condition 1: When V updraft (z)≥V terminal When (z), it means that the lifting force of the updraft is greater than the load limit settling velocity. If the fire extinguishing bomb is thrown at this point, it will not only fail to sink effectively, but will also undergo upward reverse displacement (or be carried by the airflow in the opposite direction) or hover in the air, posing a risk of dynamic penetration.

[0084] Condition 2: Preset the maximum evaporation loss rate threshold, denoted as E. threshold The preferred value is 2 kg / s;

[0085] When E rate (z)≥E threshold When the load heat loss rate exceeds the preset maximum evaporation loss rate threshold, there is a risk that it will be completely dried out or sintered and fail in mid-air before reaching the deflagration target area.

[0086] It should be noted that, in this embodiment, E threshold The overall preferred range is set at 0.5 kg / s to 3.5 kg / s, with specific configuration rules as follows: For water-based / hydrogel fire extinguishing agents, the preferred range is set at 2.0 kg / s to 3.5 kg / s. Due to its high latent heat of vaporization (approximately 2.26 × 10⁶ J / kg), it requires the absorption of significant thermal radiation to undergo a phase change, and the evaporation of water can form a vapor insulation layer on the surface, thus exhibiting the highest tolerance to instantaneous mass loss; For dry powder fire extinguishing agents (such as ammonium phosphate), the preferred range is set at 0.5 kg / s to 1.5 kg / s. Dry powder does not undergo evaporation in the traditional sense, but rather absorbs heat to undergo thermal decomposition (dehydration or release of ammonia) or melt solidification. Its phase change heat absorption is relatively small (on the order of approximately 10⁵ J / kg). If the surface agent melts prematurely due to heat, it will cause the internal agent to clump and fail, failing to form an effective cover at the deflagration point. Therefore, its fault tolerance threshold is set to be the most stringent. For special compressed air foam / gel foam, the preferred range is set to 1.0 kg / s to 2.0 kg / s. Although it contains moisture, under extreme thermal radiation impact, the outer cell structure is extremely prone to large-scale rupture and collapse due to instantaneous flash evaporation of moisture, resulting in chain failure. Its threshold needs to be set in a moderately stringent range.

[0087] The set of spatial grids that meet any of the above conditions will be uniformly marked as extreme thermal convection divergence zones (absolute no-throw zones) in the 3D topology map; and when extrapolating from top to bottom, this will be marked as the upper boundary of the divergence zone (i.e., the ultimate physical red line for safe throwing operations). Using an 80kg water-based fire extinguishing load (set E...) threshold Taking 2.0 kg / s as an example, the corresponding data calculation boundaries and judgment process are shown in Table 2. Table 2: Calculation Table of Vertical Convection and Load Settlement at Different Heights Directly Above the Fire Source (for an 80 kg spherical load)

[0088]

[0089] Through rigorous calculations in Table 2, it is found that when extrapolating downwards to a vertical height of 50m from the treetop fire source, the relative free-fall limit velocity of the 80kg heavy water-based fire extinguishing bomb (45.1m / s) will, for the first time, be unable to withstand the violent updraft at that height (46.2m / s) (first triggering condition 1). This method, through this rigorous "dynamic + thermal" dual-source calculation, accurately calibrates 50m as the "upper boundary of the divergence zone." This provides absolutely reliable data support for the subsequent requirement in Example 4 that the core operational aircraft group "must use the negative pressure funnel suction effect above the boundary of the divergence zone to perform reverse dropping, and is strictly prohibited from entering the interior of the divergence zone to perform conventional gravity airdrops."

[0090] The technical principle of this embodiment lies in the deep coupling calculation of the vertical airflow dynamics directly above the fire site with the thermodynamic loss of the fire extinguishing load. Due to the extreme heating and expansion of local air caused by ground deflagration in a massive fire, a violent vertical upward thermal convection is formed. This method uses the thermodynamic plume equation to calculate the upward airflow velocity inside the vortex tube at each altitude level based on measured temperature and altitude. Simultaneously, this method combines the fire extinguishing load's own mass and windward area with the air density at the current altitude due to high temperature, and calculates the terminal free settling velocity of the load under this environment based on the aerodynamic drag equation. Furthermore, this method introduces the blackbody radiation law to calculate the mass evaporation loss rate per unit time caused by the extreme high temperature during the load's descent. By numerically comparing the upward lifting force of the rising airflow with the load's own gravitational sinking limit, and jointly determining the evaporation loss rate with a preset safety threshold, this method can accurately delineate the boundary of the extreme thermal convection divergence zone in a three-dimensional topological map where the load cannot fall and is instead blown upwards, or even directly vaporizes and fails in mid-air. The beneficial effect of this embodiment lies in changing the traditional operational mode of blindly dropping heavy fire extinguishing bombs by gravity free fall based solely on vision or a fixed altitude. Through rigorous aerodynamic and thermal radiation physics model calculations, this scheduling method can accurately predict the actual motion of heavy fire extinguishing bombs at the center of the fire, precisely pinpointing the altitude red line that would inevitably lead to failure in conventional airdrops. This dynamic red line, derived from multi-physics field coupling calculations, prevents the core drone carrying critical payloads from futilely entering areas of turbulent updrafts, effectively preventing the fire extinguishing agent from being thrown backwards by strong winds or prematurely disintegrating due to heat. This provides a reliable altitude benchmark for subsequent core drones to remain above the boundary of the divergence zone, thereby utilizing the absolute negative pressure funnel effect at the bottom of the fire tornado vortex tube to perform reverse suction and dropping, effectively ensuring the accuracy of aerial firefighting operations and the effective utilization rate of high-value fire extinguishing agents.

[0091] Example 4:

[0092] The first asymmetric cluster penetration strategy is triggered, and an aerodynamic shielding volume is dynamically constructed. The specific steps include: obtaining the entry point and target heading of the core operation drone carrying a heavy fire-fighting payload into the manifold adsorption trap zone, and extracting the tangential wind speed vector of the thermal fire tornado at the current location; according to the wind direction of the tangential wind speed vector, scheduling a preset number of light, highly mobile escort drones to generate an asymmetric wind-breaking formation on the windward side and rear side of the core operation drone, and calculating the density coefficient on the windward side, where the spacing between the escort drones on the windward side is smaller than the spacing on the leeward side; based on the rotor downwash airflow velocity of the escort drone swarm, the physical shielding area of ​​the airframe, and combined with the wind attenuation law of the three-dimensional thermal vortex, a shielding attenuation coefficient is introduced to calculate the equivalent local low-wind-speed stable flow field three-dimensional space generated by the escort drone swarm in the space where the core operation drone is located, which is defined as the aerodynamic shielding volume; dynamically adjusting the angle of attack and three-dimensional relative position of the escort drone swarm so that the flight attitude of the core operation drone is always completely enveloped within the boundary of the aerodynamic shielding volume to forcibly penetrate the outer edge of the fire line.

[0093] Model B (core unit, carrying 80kg dry powder fire extinguishing bombs) has poor anti-disturbance capabilities; if it directly enters an area with MAI > 1.0, it will be overturned by strong winds. This method deploys six highly maneuverable escort aircraft (Model A) to form a dense physical airflow barrier at the flanks and front, utilizing the wake effect of hydrodynamics to create a dynamic three-dimensional aerodynamic safety envelope. Due to the strong asymmetry of the tangential winds of fire tornadoes, the formation must be dynamically deflected.

[0094] S301, Tangential wind field analysis and array yaw sub-steps include: real-time reading of the incoming wind direction in the current spatial coordinates of the core machine using multi-source sensors, and setting the incoming wind azimuth angle θ in the absolute coordinate system. wind =45°. The guard cluster deploys an asymmetric topological formation with the core unit as the origin (0,0,0), facing the wind at a 45° angle. The centripetal force of the fire tornado causes the wind pressure to be greater closer to the vortex center. Therefore, the side of the formation directly facing the vortex center and the prevailing wind (the windward side) requires stronger kinetic energy consumption capabilities and needs to be densely arranged; while the side away from the vortex center (the leeward side) experiences less wind force and can be sparsely arranged to save the overall power of the cluster.

[0095] Defining the density coefficient ρ on the windward side windward The principle is that the spacing between the escort aircraft on the windward side (the side directly impacted by the fire tornado's shear wind) is set to the rotor diameter of the escort aircraft. 1.2 times; specifically, the critical value constrained by the gap throttling effect of fluid dynamics, with the leeward side spacing set as the rotor diameter of the escort aircraft. 2.5 times, specifically meaning that the spacing between escort aircraft on the leeward side is widened to an independent threshold that does not cause wake interference.

[0096] S302. Based on the flow tube momentum theory and the law of conservation of mass, calculate the wind speed deficit ΔUk(x) in the space behind a single escort aircraft k when flying at a high angle of attack against the wind, caused by the rotor's work extracting kinetic energy from the air. a The formula is as follows:

[0097]

[0098] in, The real-time tangential wind speed at the fire scene is expressed in [m / s] (e.g., 30m / s). The dimension is [m], which represents the downwind longitudinal axial distance between the geometric force center of the core engine and the rotation center of the escort aircraft's rotor. Let be the rotor diameter of the escort aircraft, with dimensions [m], and a be the axial induction factor, with dimensions being a dimensionless number. The wake expansion coefficient represents the cone angle at which the airflow spreads outwards after passing over the rotor. In the extremely unstable atmospheric boundary layer characterized by high temperature and turbulence during forest fires, fluid viscosity changes, affecting the wake expansion coefficient. The safe reference temperature T of the current environment is read using an onboard infrared temperature sensor. env With the real-time core temperature T at the front of the fire fire The wake expansion coefficient is dynamically calculated using the following thermodynamic and turbulence coupling equations. :

[0099]

[0100] in, The standard laminar flow reference expansion coefficient is a dimensionless constant of 0.05, representing the normal aerodynamic dissipation rate under normal conditions without fire disturbance. For the actual measured temperature of the grid, The safe baseline atmospheric temperature outside the fire area is 25°C, which corresponds to 298K. The gain coefficient for thermal convection-turbulence distortion coupling represents the degree of turbulence diffusion intensification caused by a doubling of temperature, and is preset to 0.05. When the drone swarm advances towards the fire line, the temperature rises sharply, and the wake expansion coefficient is recalculated in real time at a frequency of 10Hz. The calculation of the windbreak array is updated within the control of this method, ensuring that it always closely matches the actual atmospheric boundary layer state in a fire situation. The specific dynamic calculation process is shown in Table 3.

[0101] Table 3: Dynamic thermal wake expansion coefficient k under different fire temperature layers w Real-time solution table

[0102]

[0103] This represents the relative distance between the core engine and the escort aircraft in front. In fluid dynamics, the decay law of the wake does not depend on the absolute distance; that is, it is the downwind axial distance between the geometric center of the core engine and the center of rotation of the escort aircraft's rotor. It depends not on the distance itself, but on the relative distance.

[0104] This leads to the distance directly behind the escort aircraft. The residual wind speed U at the location wake,k (x a ):

[0105]

[0106] When the escort aircraft tilts its fuselage significantly (increasing the angle of attack) and outputs full power to resist strong winds, the axial induction factor α increases. The calculation steps are as follows: S3021. When the UAV's rotor (propeller) rotates into the wind, it is not a completely closed "airtight solid flow-blocking surface," but rather an actuated disk with a certain air permeability. This represents the percentage by which the original ambient wind speed is "intercepted" and slowed down when passing through this rotor disk. Based on the actuated disk theory, the real-time thrust coefficient C is calculated. T The formula is as follows:

[0107]

[0108] in, This represents the horizontal wind-resistant thrust currently output by the flight control system, with dimensions in Newtons. This represents the density of the high-temperature air in the current fire environment, with dimensions [kg / m³]. 3 ]; The original tangential wind speed at the fire scene is expressed in [m / s]; The equivalent frontal sweep area of ​​the escort aircraft rotor, with dimensions [m 2 The calculation formula is: ;

[0109] S3022. Based on the conservation of flow tube momentum, the real-time thrust coefficient C T The relationship with the axial induction factor is a quadratic equation in one variable: C T =4a(1-a). The physical formula for solving this equation is: ;

[0110] It should be noted that the above quadratic equation is based on the assumption of an ideal steady-flow tube, and theoretically the thrust coefficient C... T There is a mathematical upper limit (i.e., C). T ≤1). However, under extreme gusts of fire tornadoes and strong thermal convection shear, the UAV rotor is very prone to entering a turbulent wake state. The sudden and violent change in dynamic pressure of the airflow will cause the real-time C calculated by the flight control processor to be affected. TBriefly and sharply greater than 1. Once C T >1. The value within the square root of the above formula will be converted to a negative number. The underlying processor's floating-point square root instruction will directly output an imaginary number, potentially causing a fatal arithmetic overflow and crash in the underlying flight control program, leading to the drone losing its attitude calculation ability and crashing. This method embeds a forced physical truncation algorithm: [The last part, "in real-time C," appears to be an error and doesn't translate directly. It likely refers to a specific algorithm or instruction.] T Before substituting into the quadratic equation, perform Boolean logic probing; when C is detected... T When the value is ≥0.99 (with a safety margin of 0.01 for floating-point precision to prevent overflow), C is forced to be... T A non-linear truncation is applied, forcibly locking its value to 0.99, thus 1-C T The constant is a non-negative real number. Nonlinear truncation ensures that the flight control system maintains the robustness of the solution program under extreme turbulent impacts in a fire scene, outputting a maximum limiting induction factor a = 0.4, guaranteeing that the UAV's attitude does not collapse in severe wind conditions. It should be noted that the maximum wind-resistant thrust F... max This is the static physical maximum value of the drone hardware, that is, the maximum physical thrust that the motor and propeller can generate when the motor output power reaches its peak; the horizontal wind-resistant thrust currently output by the flight control method. Treq , represents the dynamic real-time demand value (current real-time thrust output command) of the UAV when performing a mission, and is a variable that constantly changes with wind speed, and T req ≤F max Real-time thrust coefficient C T The dimensionless hydrodynamic ratio, representing the thrust T generated by the UAV, is a measure of the dynamic force. req The proportional relationship between the "natural wind dynamic pressure passing through the rotor area" and the "natural wind dynamic pressure".

[0111] S303. Substituting the single-machine physical equations of S302 into the asymmetric coordinate system of S301, the final "eccentric / asymmetric" overall shielding flow field is calculated. In the asymmetric formation defined by S301, the core operational UAV is simultaneously located in the cross wake region of six escort aircraft. Due to the close proximity and large angle of attack (large 'a' value) of the windward escort aircraft, and the greater distance and smaller angle of attack (smaller 'a' value) of the leeward escort aircraft, the final superimposed flow field is an asymmetric flow field strongly biased towards the windward side under pressure.

[0112] According to the principle of conservation of kinetic energy in the flow field, the wind speed deficits caused by these six obstruction sources located in asymmetric coordinates cannot be linearly added together, but rather are a linear superposition of their kinetic energy deficit terms (the square of the velocity). Therefore, the coordinates of the core machine location point (X...) c ,Y c Z c The combined attenuation of wind speed U at point ) core,pos The solution formula is:

[0113]

[0114] in, It is the wind speed deficit generated by the kth escort aircraft at the core coordinate point in the asymmetric formation, calculated according to the S302 formula, with the dimension [m / s]. The coefficient is set as the fire thermal dynamics correction factor. Due to the asymmetrical and dense arrangement of multiple drones, the vortex interference at the rotor tips on the windward side is much more intense than on the leeward side, and the high temperature gradient at the fire site can cause nonlinear "air leakage" in the windbreak wall. K is set as follows: shield =0.82, as a safety margin correction, deducting the energy loss caused by multi-machine interference;

[0115] Step S303 calculates the wind speed at the core engine's "center point" through multi-source superposition. However, in actual engineering, the core aircraft (model B) is not a geometric point mass without volume, but a three-dimensional entity with a full load of 120kg and a huge wingspan. More importantly, due to the "asymmetric" array arrangement of S301, the low-wind-speed safety zone generated by the superposition of multiple aircraft wakes appears in spatial geometry as an eccentric, irregular, and locally concave three-dimensional fluid cavity. If only the safety of the core engine's "center point" is guaranteed, its rotor tips or outer arms are very likely to have penetrated the irregular wake boundary, directly exposed to the deadly high-speed shear wind, instantly triggering "asymmetric roll" and causing a crash. To calculate the wind speed at a single point, it is necessary to extract the safety isosurface boundary of the entire asymmetric flow field in the three-dimensional mesh and forcibly constrain the entire physical shape envelope of the core engine within this boundary. Therefore, it is necessary to construct an asymmetric three-dimensional aerodynamic shielding envelope.

[0116] S304, the sub-step for extracting the spatial boundary and maintaining the geometric constraint of the asymmetric three-dimensional aerodynamic shielding envelope, and the preset limit safety wind speed threshold V of the core machine. safe The preferred value is set to 15.0 m / s (meaning that the probability of lateral overturning is greater than 80% when the wind speed exceeds this). The wind speed calculation is then performed from the center point (X). c ,Y c Z c Extending to the three-dimensional local coordinate system domain Ω behind the guard formation local Internally, the wind speed after comprehensive attenuation is calculated by traversing all grid points (x, y, z) in space. Define Not exceeding the extreme safe wind speed threshold V safe Asymmetric aerodynamic shielding volume Its mathematical expression is:

[0117]

[0118] At the same time, extract the solid-state 3D physical body point cloud set E of model B itself. core(This set includes the maximum three-dimensional geometric boundary swept by the UAV fuselage, outwardly extending arms, and high-speed rotating rotor). During the penetration of fire tornadoes, this method performs real-time spatial determination at a frequency of 100Hz, establishing the core constraint as a topological inclusion relationship:

[0119]

[0120] That is, every three-dimensional coordinate point within the geometric envelope of the core machine must completely belong to the safe envelope set. Once a gust of wind in the fire causes... The boundary is contracting inward, and detection is made. This triggers the first asymmetric cluster penetration strategy, including: prioritizing the control of the flanking escort aircraft to increase the current pitch angle of attack by 15% to 25%. While increasing the angle of attack, this method simultaneously shortens the downwind axial distance x between the windward escort aircraft and the core engine by 10% to 15%. a until it forcibly penetrates the outer perimeter of the fire line.

[0121] The principle of attitude priority compensation (increasing angle of attack) is to prioritize controlling the side escort aircraft to increase the current pitch angle of attack by 15% to 25% (the absolute maximum angle of attack is limited to no more than 45° to prevent loss of lift due to rotor stall). This is based on the actuation disk theory in S302. T The quadratic equation with variable a, and the surge in angle of attack within this interval, can increase the thrust coefficient C. T This generates a nonlinear jump, which rapidly brings the axial induction factor a close to the theoretical limit (increasing it to between 0.35 and 0.45), instantly acquiring more air momentum to impede the incoming wind.

[0122] The principle of spatial compensation (reducing depth distance) is that, while increasing the angle of attack, this method simultaneously reduces the axial distance in the downwind direction by 10% to 15% between the escort aircraft on the windward side and the nuclear aircraft. Due to the wake attenuation term, i.e., S302, the formula... The axial distance in the mid-to-downwind direction is quadratically sensitive; shortening this distance by 10%–15% can result in an equivalent gain of approximately 20%–30% in shielding wind speed. This reduction is limited by the escort aircraft's collision safety threshold; the escort aircraft's collision safety threshold is set at a minimum relative distance ≥ 1.5 × D. rotor ;

[0123] In this embodiment, the technical principle lies in utilizing the rotor wake interference effect in fluid mechanics to dynamically construct a windproof barrier for the core aircraft. This method analyzes the intense asymmetric wind shear of the fire tornado and densely deploys highly mobile escort aircraft on the windward side of the core aircraft to form an asymmetric wind-breaking formation. Based on flow tube momentum theory, this method calculates the wind speed deficit generated after the escort aircraft's rotors extract aerodynamic energy at high angles of attack, and introduces the wake expansion coefficient caused by the high temperature of the fire scene. Combining the principle of multi-aircraft energy superposition, it deduces a low-speed aerodynamic shielding envelope biased towards the windward side for pressure resistance. This method extracts the three-dimensional physical entity point cloud of the core aircraft, forcing all fuselage boundaries to be strictly enveloped within this safe fluid cavity. During the penetration process, this method frequently monitors topological inclusion relationships. If a sudden gust of wind causes the shielding boundary to shrink, it immediately controls the windward escort aircraft to increase its pitch angle of attack to extract more aerodynamic momentum and simultaneously shortens the downwind depth distance from the core aircraft, instantly obtaining windproof equivalent gain using the wake attenuation law. The beneficial effect of this embodiment lies in solving the problem of heavy-payload UAVs having poor anti-disturbance capabilities and being easily overturned and crashed by shear winds. The asymmetric formation breaks away from the traditional uniform arrangement mode, concentrating windproof resources on the dangerous windward side, while saving the overall power of the cluster. The constraint condition of fully enclosing the three-dimensional physical boundary within the fluid safety isosurface completely avoids asymmetric roll loss of control caused by single-point wind speed meeting the standard but wing edges being exposed. The introduction of a dual attitude and spatial compensation mechanism based on momentum and thermodynamic coupling enables the cluster to actively strengthen the thickness of the windbreak when encountering gusts. This leap from passive obstacle avoidance to active flow field reshaping ensures that the heterogeneous cluster maintains the robustness of its topology in an extremely unstable thermodynamic boundary layer, ensuring that the core operational UAV can forcefully penetrate the deadly outer barrier.

[0124] Example 5

[0125] Example 4 mainly occurs in the "manifold adsorption trap zone" (because here the main force is horizontal tangential strong wind and centripetal suction, and the drones will be "sucked in" or "blown over"); the operation steps of the drone swarm to perform the mission are: take off from the periphery → form a formation to penetrate the trap zone → replace the damaged units → finally arrive at the divergence zone and drop bombs; in the obvious characteristics of the extreme thermal convection divergence zone, the hydrodynamic characteristics change drastically: from "horizontal shear wind" to "vertical violent updraft".

[0126] The first operational phase, specifically the lateral penetration phase, involves the UAV swarm traversing the manifold adsorption trap zone. The primary hydrodynamic threats are strong horizontal shear wind pressure and centripetal negative pressure. The escort aircraft (Type A) at the forefront of the asymmetric formation on the windward side bears the greatest thermal convection impact. If a particular escort aircraft rapidly depletes its power or encounters extremely strong thermal turbulence, and adopts the conventional UAV logic of "headwind deceleration, hovering, or reverse escape," it will instantly be sucked into the vortex core by the centripetal negative pressure and crash due to insufficient thrust. To address the spatial properties of the manifold adsorption trap zone, a second strategy is triggered: a tangential slingshot escape and swarm formation self-healing. This includes S401 to S403.

[0127] S401. Disrupt Conventional Flight Control and Calculate the Centrifugal Escape Critical Value: Real-time monitoring of the manifold adsorption index (MAI) of each UAV in the formation. When any escort UAV loses cluster aerodynamic shielding due to strong thermal turbulence, or when its own MAI rises sharply and exceeds the preset limit threshold (e.g., MAI=0.95, close to the 1.0 boundary), this method immediately hard-disrupts the conventional flight control's headwind deceleration avoidance and reverse hovering escape commands. This method obtains the tangential wind speed vector of the thermal fire tornado field at the current coordinates of the UAV. The radial distance *r* relative to the center of the vortex. Calculate the centrifugal escape critical value (velocity threshold) *V* by combining the UAV's own mass *m* with the current remaining maximum instantaneous thrust output. critical The physical principle behind this is to utilize the enormous centrifugal force generated by the aircraft's acceleration with the wind to counteract the centripetal negative pressure suction force F of the fire. suction When the centrifugal force is greater than the centripetal force, escape can be achieved. The calculation equation is as follows:

[0128]

[0129] In this equation, the left side represents the outward centrifugal force generated by the drone's circular / spiral motion; the right side represents the centripetal negative pressure suction acting on the drone's fuselage. According to the divergence theorem in fluid mechanics, the centripetal negative pressure suction is equal to the centripetal negative pressure gradient calculated by S202. Equivalent closed volume of the drone The product of, i.e.:

[0130]

[0131] It should be noted that under extreme aerodynamic conditions, V env It is not obtained by dividing the drone's mass by the ambient air density (this algorithm would calculate the volume of the expanded hot air instead of the drone's actual volume, leading to severe distortion). In this embodiment, V envThe fixed factory-preset constants for flight control are numerically equal to the three-dimensional CAD watertight external geometric volume formed by the UAV's core fuselage, arm configuration, and external firefighting payload. For example, for a quadcopter escort UAV with a weight of 5kg, the equivalent closed external volume V of its core fuselage and components is... env Calibrated at 0.015m 3 ~0.025m 3 Between (approximately equal to the actual volume occupied by a miniature storage box);

[0132] Substituting the suction formula into the equilibrium equation, the minimum tangential composite velocity (centrifugal escape critical value) required to prevent the drone from being sucked into the core of the fire is obtained as follows:

[0133]

[0134] S402. Execute the tangential smooth acceleration and large-angle jettison (slingshot release) strategy to control the endangered UAV to release its headwind and pressure resistance attitude, adapt to the environmental flow field, and follow the tangential wind speed vector. Applying maximum positive thrust in the positive direction generates tangential smooth acceleration, increasing the fuselage's absolute velocity. Tangential wind speed at the fire site By performing forward vector superposition, the composite velocity of the UAV in the fire vortex is obtained. Real-time monitoring of the composite velocity of the drone within the fire vortex; When detected This indicates that the ratio of the square of the combined velocity to the radius of the vortex center satisfies the centrifugal escape critical value, triggering a slingshot-like ejection maneuver. The thrust vector ejection angle α is calculated based on the topological curvature of the three-dimensional thermodynamic flow field. escape The thrust direction is changed from purely tangential to an outward (away from the vortex center) angle of 15 to 30 degrees (preferably 22.5 degrees, which generates sufficient radial escape component without causing tailwind sideslip stall). The UAV is controlled to output thrust at full power according to the thrust vector throw angle. Utilizing the extremely powerful wind and centrifugal force field of the thermal fire tornado itself, the UAV is instantly thrown out of the manifold adsorption trap zone boundary along a parabolic trajectory, thus achieving self-preservation.

[0135] The thrust vector ejection angle α is calculated based on the topological curvature of the three-dimensional thermodynamic flow field. escape The specific technical steps are as follows:

[0136] S4021, When the drone is at a high synthetic motion speed When rotating at high speed with the shear wind at a fire scene, if the thrust is directly deflected outward by 0° (i.e., flying outward perpendicular to the wind direction), the extremely high dynamic pressure shear wind will directly impact the side of the UAV's fuselage laterally. This will instantly trigger a severe "lateral aerodynamic sideslip stall," causing the UAV to tumble and crash. Therefore, this method presupposes the maximum critical sideslip angle α based on the aircraft's aerodynamic shape. max α serves as the absolute physical upper limit of the escape deflection angle. To adapt to different configurations of operational clusters in fire rescue, this method uses α... max The overall preferred range is set to 10°~45°, and the specific adaptive configuration and physical constraint rules are as follows: For large work aircraft / coaxial twin-rotor aircraft carrying heavy fire extinguishing bombs (such as those carrying an 80kg payload), α max The preferred range is set at 10°~20° (most stringent). Due to the large physical load on the chassis, this type of aircraft has a large lateral windward cross-sectional area and a large distance between its center of gravity and the crosswind aerodynamic pressure center, exhibiting a strong "sailboat effect." Even a tiny change in crosswind can generate fatal roll torque; therefore, the nose must be firmly aligned with the tangential gale (maintaining a very small sideslip angle), relying on a very small outward component force to slowly cut out the manifold boundary like slicing a pancake, resulting in a low margin for error. For a conventional medium-sized escort quadcopter / hexacopter UAV (model A in this embodiment), α max The preferred range is set between 20° and 35° (in this embodiment, the preferred absolute boundary value is 30°). Physical basis: This type of aircraft (core fuselage outer volume approximately 0.015m³) 3 ~0.025m 3 The box-type fuselage structure has moderate lateral drag. Wind tunnel data and fluid simulations show that when the crosswind angle exceeds 30°, the airflow separation boundary generated on the side of the fuselage expands sharply, just enough to engulf the leeward rotor's plane of rotation. Therefore, 30° is the final physical red line to ensure that the leeward rotor does not stall due to "flow field stripping". For streamlined / racing forward reconnaissance UAVs (FPV configurations), α max The preferred range is set between 35° and 45°. Physical basis: These aircraft types employ a teardrop or bullet-shaped fairing design, resulting in low lateral drag and allowing airflow to easily remain attached to the sides of the fuselage without boundary layer separation. Therefore, they possess extremely strong resistance to sideslip and crosswinds, allowing for rapid yaw cuts out of danger zones at larger, more aggressive angles (up to 45°).

[0137] In the purely tangential flow field of a thermal fire tornado, the local topological curvature K of the streamlines is... topo It is strictly inversely proportional to the radius r of its distance from the vortex center, i.e., K topo=1 / r. The closer the UAV is to the core region, the more drastically the streamlines curve inward (the greater the curvature), and the steeper the required outward escape angle; conversely, the closer to the outer edge (the smaller the curvature), the gentler the escape angle. This method extracts the vortex core radius R calibrated in S201. core (As the reciprocal of the maximum curvature of the reference), combined with the real-time centripetal radius r of the current location of the UAV, the precise thrust vector jettison angle α is dynamically calculated using the following trigonometric function nonlinear mapping formula. escape :

[0138]

[0139] This formula describes the spatial geometric topology of the flow field. This is converted into flight control attitude commands for the drone. An example is:

[0140] When in a perilous situation, if the drone is sucked to the edge close to the fatal core area (i.e., r≈Rcorer, meaning the distance is twice the distance), then... The formula calculates α escape =30°×sin(90°)=30°. This method outputs the maximum limiting deflection angle, forcibly breaking through the constraints of a high-curvature streamline with the most powerful radial thrust. When outside the trap, if the UAV triggers escape at a distant manifold edge (e.g., r=2R), corer When the distance is twice the normal distance, the formula calculates α. escape =30°×sin(45°)≈21.2°. This method outputs a moderate deflection angle, allowing it to slide smoothly out. When within the safe boundary, if r=3R corer , representing a distance of three times, calculate α. escape =30°×sin(30°)=15°; Through dynamic calculation using three examples, the thrust direction is changed from pure tangential to an outward deflection by an exact α. escape The included angle (the output range is strictly and continuously constrained between 15° and 30°). The drone is controlled to output thrust at full power at this angle, utilizing the extremely powerful wind and centrifugal force field of the thermal fire tornado itself to instantly throw the drone out of the manifold adsorption trap zone boundary along a parabolic trajectory, thus achieving self-protection.

[0141] S403. Calculate the dynamic manifold replacement rate and redundancy maneuvering:

[0142] When the outermost escort drone on the windward side triggers the aforementioned slingshot escape trajectory and detaches from the formation, a spatial topological void will momentarily appear in front of the core drone (i.e., a physical gap in the aerodynamic shielding formation's spatial topology). The three-dimensional coordinates of this detachment node are extracted. Based on Bernoulli's Orifice-Flow Principle, the process of high-pressure airflow from outside the formation jetting into the low-pressure wake region inside the formation is equivalent to a pressure-difference-driven orifice jet model. The penetration and damage rate of the high-temperature fluid from the external fire scene jetting inward through this void is then calculated. The calculation equation is as follows:

[0143]

[0144] in, The high-temperature local air density of the current spatial grid is calculated in real time for step S202.

[0145] To protect the outside of the formation, the fire tornado shear wind field directly impacts the total pressure of the fluid stagnation point on the windward side of the formation; The local hydrostatic pressure within the asymmetric aerodynamic shielding envelope where the core machine built for step S304 is located; the huge pressure difference between the two. It is the driving force that propels the thermal fluid to penetrate inward; The equivalent orifice contraction coefficient is given by the physical gap in the formation. Considering that the fluid is forced through the irregular gaps formed by the arms and high-speed rotating rotors of the adjacent escort drones on both sides, the fluid boundary layer will generate strong viscous friction and eddy current dissipation. In this embodiment, C... gap The preferred value is set to 0.61 to 0.65;

[0146] After calculating the actual rate of infiltration failure Subsequently, this method calculates the dynamic manifold replacement rate and the limiting safety time window T. window :

[0147]

[0148] Among them, D gap K represents the physical width of the gap in the formation. safety The thermal expansion safety factor is taken as 1.2.

[0149] When extremely high-temperature fluids (e.g., above 800°C) from the fire scene breach the physical gaps in the windbreak wall and invade the low-pressure cold zone, they not only possess an initial penetration rate... It will also experience extremely violent volume expansion and convective acceleration (i.e., thermodynamic shock wave effect) due to the huge temperature gradient. This nonlinear expansion and acceleration means that the time it takes for the leading edge of the real thermal fluid to contact the rear core is necessarily shorter than the theoretical time calculated by ideal uniform division; K safety The preferred range is strictly defined between 1.10 and 1.50. The lower limit of 1.10 is used when the temperature difference between the inside and outside of the fire is small (e.g., in the outer area); the upper limit of 1.50 is used when encountering an extreme deflagration center. In the specific fire tornado penetration operation of this embodiment (the coordination distance between the Type A escort drone group and the 80kg core drone is 2 to 3 meters), the preferred value is 1.2. This value is the optimal energy game solution, ensuring sufficient lead time (leaving a 20% time margin to guard against sudden thermal turbulence changes) for launching fully loaded backup drones for attack, while minimizing the waste of cluster redundant kinetic energy caused by frequent false alarms of "false backups" due to excessively large coefficient settings. In T window Within a very short window (typically 0.5 to 1.5 seconds), this method immediately dispatches a backup escort UAV in a safe redundancy state from the leeward side (low-pressure area) of the asymmetric formation. Because the leeward UAV was sparsely deployed and subjected to low wind pressure during the initial execution of step S301, it retains sufficient power and maneuverability. This method controls the backup UAV to operate at its maximum maneuverability V. max It crosses above or behind the core unit, cutting diagonally and precisely filling the spatial topological gaps on the windward side. Through this dynamic replacement mechanism, it ensures that the aerodynamic shielding volume Vs,wake outside the core operation UAV does not suffer fatal physical damage in the event of single-unit combat damage / escape, ensuring that the core unit can ultimately and accurately drop the 80kg dry powder fire extinguishing bomb at the ignition point at the bottom of the vortex.

[0150] The second operational phase, specifically the vertical suppression phase, occurs after the second tangential slingshot escape and cluster formation self-healing strategies (S401 to S403) are executed. Under the continuous aerodynamic shielding of the dynamic formation replacement, the core UAV successfully penetrates the boundary of the manifold adsorption trap zone laterally. At this point, this method, using a 3D environmental topology map and onboard barometers, determines that the core UAV has reached directly above the 3D vortex core area, officially entering the extreme thermal convection divergence zone. After crossing the spatial boundary, the hydrodynamic characteristics of the fire area undergo a transient and dramatic change: from predominantly high-dynamic-pressure horizontal tangential winds to vertical extreme thermal convection updrafts triggered by ground deflagration. Therefore, this method automatically interrupts the windbreak logic of the first operational phase and switches to a negative-pressure delivery step based on vertical convection field analysis.

[0151] S404, the negative pressure throwing steps across the manifold trap boundary and based on the extreme thermal convection divergence region include:

[0152] After the core operational UAV successfully penetrates the manifold adsorption trap zone under the continuous aerodynamic shielding of the dynamic formation backup, this method uses a 3D environmental topology map and onboard barometers to determine that the core UAV has reached directly above the 3D vortex core area, thus officially entering the extreme thermal convection divergence zone. Within this zone, the hydrodynamic characteristics of the fire site undergo a transient and dramatic change: from predominantly high-dynamic-pressure horizontal tangential winds to vertical extreme thermal convection updrafts (longitudinal wind speed V) triggered by ground deflagration. updraft (Rapidly rising). Perform the following firefighting procedures:

[0153] Step S4041, aerodynamic dispersal and updraft-riding flight of the escort swarm: After entering the divergence zone, horizontal windbreaks are no longer needed. This method dismantles the asymmetric wind-breaking formation, and the remaining escort UAVs quickly switch to a "wave-riding" attitude, using the powerful updraft to provide lift to save energy, and providing surveillance around the core aircraft.

[0154] Step S4042, Vertical Convection Field Analysis and Load Detonation Height Calculation: The core operation UAV uses a bottom-mounted downward-looking multispectral radar to read the vertical updraft velocity V in the core area of ​​the three-dimensional vortex tube directly below. updraft The central negative pressure value. This method does not use the traditional "gravity free fall" throwing method, because under extreme updrafts, an 80kg dry powder fire extinguishing bomb would be blown off course by strong winds or even blown back into the airspace along the same path.

[0155] Step S4043, the negative pressure suction and throwing step, extracts the predicted value V of the fire extinguishing projectile's aerodynamic terminal free settling velocity calculated at the calibration boundary of the extreme thermal convection divergence zone in the aforementioned Example 2 (Step S206) using this method. terminal and real-time vertical ascent speed V updraft Only when V terminal >V updraft Only then is the release lock released. More importantly, combining the absolute negative pressure funnel effect that inevitably exists at the center of a vortex in fluid mechanics, this flight control method controls the core engine to precisely release the vortex at the geometric center of the thermal convection divergence zone (i.e., the vortex tube eye), and sets a delayed fuse to detonate at the throat of the vortex tube (the phase change section where the airflow changes from rotation to straight up). Specifically, this includes:

[0156] Based on the three-dimensional voxel grid network constructed in Example 2, the system extracts the tangential wind speed of the fire at each height level along the vertical axis (Z-axis) of the fire tornado's eye center downwards. With vertical updraft speed In subsurface fluid dynamics, the upper throat of a fire tornado is the critical layer where the fluid kinetic energy profile undergoes a sudden change. This method calculates the swirling kinetic energy ratio Sa(z) along this vertical axis using the following equation:

[0157]

[0158] Scan the grid from top to bottom. When a reversal of Sa(z) value is detected, crossing the threshold of 1.0 (i.e., from the top...), the grid is scanned. The mutation is as follows At that time, the coordinate height is precisely locked to the phase transition section Z at the throat of the vortex tube. throat (Unit: meters). Below this cross-section is the funnel inlet where the core of the fire tornado experiences the strongest rotating suction force and the most extreme negative pressure;

[0159] Because the fire extinguishing bomb traverses ΔZ=Z during its descent. drop -Z throat When the spatial distance is such that the air density ρ it faces is... fire (z) and real-time vertical ascent velocity V updraft (z) varies drastically and nonlinearly with altitude. This method extracts the predicted terminal free settling velocity V of the fire extinguishing bomb at each altitude level calculated in Example 2 (step S206). terminal (z); The current drop height Z of the drone drop To the target detonation height Z throa Using a three-dimensional grid between points t as the integration domain, and based on the relative net velocity during the descent (downward settling limit minus upward airflow lift), a discretized dynamic path integral is performed to calculate the absolute physical time T required for the fire extinguishing projectile to reach the target cross-section. fuse :

[0160]

[0161] To ensure hardware redundancy and prevent premature fuse failure due to extreme heat radiation, the system performs microsecond-level corrections to this time and then programs the delay in milliseconds into the electronic fuse of the 80kg dry powder fire extinguishing bomb via the airborne fire control bus; this delay is implemented only when the current altitude is confirmed. terminal >V updraf (To ensure the fire extinguishing grenade has downward net penetrating power) the flight control system's core engine aligns with the vortex tube eye and releases the release lock. The fire extinguishing grenade precisely exhausts its T-force during free fall. fuse After the set delay, at Z throa T-phase cross-section initiation. Through the above steps, the large mass of dry powder extinguishing agent after detonation will not be dispersed by the rising airflow. Instead, it will be strongly drawn downward to the core deflagration zone at the bottom by the "negative pressure funnel" of the fire tornado itself, just like the absolute negative pressure funnel effect at the bottom of the vortex. The extinguishing agent will be 100% dispersed and covered by the thermodynamic wind field of the fire site, thus completely completing the task of suppressing the fire tornado.

[0162] The technical principles of this embodiment encompass the cluster formation self-healing mechanism during the lateral penetration phase and the favorable negative pressure throwing mechanism during the vertical suppression phase. During lateral penetration, when the escort aircraft on the windward side is nearing the loss of control limit due to extreme flow field impact, this method hard-blocks the conventional flight control's deceleration and hovering avoidance logic. This method calculates the critical value of the centrifugal escape velocity by extracting the real-time centripetal negative pressure gradient and tangential wind speed, and dynamically calculates the outward deflection thrust jettison angle by combining the topological curvature of the streamlines changing with radius in the surrounding space. The endangered single aircraft is controlled to output full-power thrust along the shear wind direction, utilizing the powerful centrifugal force of the fire field's wind field to instantly throw it out of the trap zone along a parabolic trajectory. Simultaneously, this method calculates the thermal fluid infiltration damage rate and the critical time window after the formation is damaged, and quickly dispatches a backup aircraft from the leeward side with lower wind pressure to obliquely cut in and fill the gap, maintaining the integrity of the aerodynamic shielding volume surrounding the core aircraft. After entering the vertical suppression phase directly above the fire source, this method disbands the windproof formation. Under the premise of confirming that the settling velocity at the end of the fire extinguishing bomb is greater than the vertical upward airflow velocity, the core machine is controlled to precisely deliver the load to the center of the vortex tube by utilizing the absolute negative pressure funnel effect that inevitably exists in the center eye of the fire tornado, so that the detonated agent is reverse-drawn to the bottom core deflagration zone.

[0163] The beneficial effects of this embodiment lie in providing a complete closed-loop solution for survival and operation in harsh flow fields. The innovative tangential slingshot escape strategy breaks away from the conventional passive approach to countering high-dynamic-pressure wind fields, cleverly utilizing the inherent kinetic energy (centrifugal force) of the flow field in the disaster-causing environment to achieve efficient self-rescue for a single aircraft, avoiding unnecessary power consumption and crashes. The dynamic redundancy replacement mechanism ensures that the heterogeneous cluster can achieve instantaneous self-healing of its physical topology when facing localized damage, promoting the absolute safety of the core aircraft. In the final delivery stage, this method abandons the traditional gravity-based free-fall blind delivery, cleverly utilizing the negative pressure physical characteristics inside the fire tornado for suction delivery. This not only ensures that heavy extinguishing agents can accurately reach the fire's throat area despite the interference of violent updrafts, but also utilizes the cyclone field itself to achieve comprehensive dispersion and coverage of the agents, effectively improving the success rate of aerial firefighting in complex forest fire environments.

[0164] like Figures 1 to 4 As shown, Figure 1This image shows the basic topology of the UAV swarm during its lateral penetration of the fire field in the first operational phase. A set of thick solid arrows in the upper left corner represents the strong tangential wind speed vector and centripetal negative pressure gradient induced by the fire. The large cross-shaped multi-rotor UAV in the center of the image is the core operational UAV. In response to this unidirectional wind shear, the system triggers the first asymmetric swarm penetration strategy: on the windward side facing the incoming wind, multiple small escort UAVs are deployed in a high-density arc array; on the leeward side, they are deployed in a low-density sparse array. The eccentric dashed boundary surrounding the core UAV precisely defines the asymmetric three-dimensional aerodynamic shielding envelope generated by the rotor wake interference of the windward UAV swarm; the physical boundary of the core UAV is completely constrained within this safe isosurface. A solid arc-shaped arrow deflected outwards in the lower left corner indicates that when the manifold adsorption index (MAI) of the escort UAV at the forefront on the windward side exceeds the ultimate bearing threshold, the system immediately disables its conventional avoidance flight control, triggering the initial jettison action of the second tangential slingshot escape strategy.

[0165] Figure 2 The focus is on demonstrating the dynamic impact of a highly turbulent and extremely unstable atmospheric boundary layer at a fire scene on the cluster formation. Figure 2 The solid arrows interspersed on the left and right sides of the escort aircraft group on the windward side represent the high-temperature fluid in the fire field pressing into the aerodynamic shielding boundary at an extremely high rate of penetration and destruction. The thick solid curved arrows on the outer side further demonstrate the trajectory of the endangered escort aircraft as it follows the centrifugal force field of the wind field and executes its separation along a specific thrust vector jettison angle. At the same time, in response to the risk of aerodynamic shielding rupture caused by the impending battle-damaged separation of the front-line aircraft, the dashed arrows on the right side of the figure fully reveal the flight path planning logic of the "dynamic formation replacement": after calculating the extreme safety time window, the system immediately activates the backup UAVs located in a safe redundancy state on the right rear of the formation (the leeward side weak pressure zone), generating a topology compensation trajectory that cuts diagonally into the windproof array on the side of the core aircraft.

[0166] Figure 3 This diagram illustrates the intermediate state between single-unit slingshot escape and spatial topology reconstruction. The parallel wind vector on the left continuously impacts the windbreak wall, while the solid parabola in the lower left corner clearly represents the critical escort aircraft's self-protective maneuver of being completely ejected from the manifold adsorption trap zone using the centrifugal force field of the fire tornado field at a preset maximum critical sideslip angle. Simultaneously, the dashed aerodynamic shielding envelope in the diagram shows a contraction trend due to the shortage on the windward side. The bifurcated dashed lines on the right, extending from the backup aircraft on the leeward side, represent the system executing a real-time spatial topology inclusion determination algorithm at high frequency (e.g., 100Hz), using the center coordinates of the core operational UAV as the origin. By calculating the three-dimensional coordinates of the spatial topology gaps left by the detached nodes, the system is verifying and selecting the optimal redundant maneuver to fill the gaps, preventing the core aircraft from being exposed to high-speed shear winds.

[0167] Figure 4This diagram provides the most intuitive global representation of the cluster formation's self-healing strategy during the first operational phase. The core operational UAVs are superimposed with orthogonal coordinate axes to define the local coordinate system. A clear "spatial topological gap" (i.e., a hole in the windbreak) is visible in the windward UAV cluster (slightly to the upper left); the large solid parabola in the lower left confirms the centrifugal escape trajectory of the escort UAV in its original position. The most crucial invention in this diagram lies in the dashed arrow running across the entire area on the right: within the extremely short time window defined by the dynamic manifold replacement rate, a backup escort UAV located in the leeward safety zone is crossing behind the core UAV at maximum maneuver speed, forcibly cutting in diagonally along this dashed trajectory, its destination precisely pointing to the topological gap in the windward formation. Through this closed-loop action, the system successfully maintains the physical integrity of the asymmetric aerodynamic shielding envelope surrounding the core operational UAVs, ensuring the continuous penetration of the UAV cluster into the core combustion zone.

[0168] Figure 5 The left side shows the operational posture of a swarm of drones performing lateral penetration in an extremely unstable thermal boundary layer: Figure 5 The diagram shows a "core operational drone" carrying a heavy firefighting payload at the center, with multiple "escort drones" densely arranged on its windward side, dynamically constructing an "asymmetric aerodynamic shielding volume," corresponding to the recorded "first asymmetric swarm penetration strategy." The diagram shows an outward tangential trajectory outside the formation, representing the "single-aircraft tangential slingshot escape trajectory" triggered by the windward escort drones due to the manifold adsorption index exceeding the ultimate bearing threshold. At the same time, a diagonal insertion trajectory behind the formation indicates that the system is scheduling backup drones to fill the gap, corresponding to the recorded "second tangential slingshot escape and swarm formation self-healing strategy." The technical roadmap comprises four rectangular flowcharts arranged downwards, corresponding to the following steps: The first flowchart, "Multi-source Data Acquisition and Hazard Zone Marking," involves acquiring meteorological and thermal field data of the fire scene and marking manifold adsorption trap zones and extreme thermal convection divergence zones in a 3D environmental topology map; the second flowchart, "Asymmetric Aerodynamic Shielding Cluster Penetration," involves scheduling multiple escort aircraft to create windbreaks around the core aircraft to execute flight routes close to the fire source; the third flowchart, "Single-Aircraft Slingshot Escape and Dynamic Replacement," involves calculating the adsorption index in real time during lateral penetration and executing single-aircraft detachment and backup aircraft formation replacement actions when the formation is in danger to maintain the integrity of the formation topology; the fourth flowchart, "Negative Pressure Suction and Throwing Suppression," involves dismantling the cluster formation after the core aircraft crosses the boundary and arrives at the divergence zone, and using the absolute negative pressure funnel effect inside the fire tornado vortex tube to execute the reverse throwing action of the extinguishing agent based on the coupled calculation of vertical convection velocity and load settling threshold value.

[0169] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, min-max-normalization and Z-score standardization.

[0170] The algorithm of this invention is implemented as a Python script. Before executing the core logic, the program first executes a data loading module (e.g., using the widely used pandas library in Python) configured to read the aforementioned spreadsheet file and load its contents into the program's working memory (e.g., a DataFrame data structure). Subsequent algorithm steps will directly query and retrieve the required configuration parameters from this in-memory data structure.

[0171] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A collaborative scheduling method for low-altitude airspace UAV swarms based on multi-source data fusion, characterized in that, The specific steps include: The system acquires a three-dimensional environmental topology map of the low-altitude airspace of forest fires, dynamic parameters of each aircraft in the UAV swarm performing firefighting and reconnaissance missions, a sequence of coordinated firefighting operations, and real-time environmental meteorological and combustion thermal field distribution data of the fire based on multi-source sensor fusion. Based on real-time environmental meteorological and combustion thermal field distribution data of the fire site, the three-dimensional vortex core area, manifold adsorption trap area and extreme thermal convection divergence area of ​​the thermal fire tornado induced by the forest fire are marked in the three-dimensional environmental topology map through three-dimensional fluid dynamics and combustion thermodynamics spatial field modeling. Obtain the initial coordinated flight trajectory of the drone swarm and determine whether each node in the flight trajectory coincides with the manifold adsorption trap region; For conventional reconnaissance aircraft whose trajectories do not overlap, standard spatial obstacle avoidance path planning should be implemented to bypass the line of fire; For the core operational drone that is required to pass through or approach the manifold adsorption trap area to deliver fire extinguishing payload, the first asymmetric swarm penetration strategy is triggered, and multiple escort drones are dispatched to dynamically build an aerodynamic shielding volume around the core operational drone. During the lateral penetration process in the first operational phase, the manifold adsorption index of each UAV in the formation is calculated in real time. When the manifold adsorption index of the windward escort UAV is higher than the preset limit bearing threshold, the second tangential slingshot escape and cluster formation self-healing strategy is triggered. The single UAV executes the tangential slingshot escape trajectory to throw itself out of the fire hazard zone, and the backup UAV is simultaneously dispatched to replace the formation. In the vertical suppression operation of the second phase of the operation, when the core operation UAV crosses the boundary of the manifold adsorption trap area and reaches the extreme thermal convection divergence area, the escort swarm formation is disarmed and the follow-up negative pressure suction and throwing based on vertical airflow analysis is carried out.

2. The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion according to claim 1, characterized in that: The specific steps for calibrating the manifold adsorption trap zone include: identifying the convergence point of the highest temperature gradient in the fire as the center coordinate of the wind field; calculating the tangential wind speed of each spatial grid around the vortex core area based on the principle of conservation of angular momentum in fluid dynamics; marking the three-dimensional spatial grids that are less than the radius of the vortex core at the center coordinates and whose central airflow is in a strong spiral upward state as the three-dimensional vortex tube core area, which is regarded as an absolute no-fly zone; for the spatial grids around the three-dimensional vortex tube core area, calculating the manifold adsorption index based on the local air density attenuation caused by the extremely high temperature in the fire, the centripetal negative pressure gradient, and the upper limit of the maximum wind resistance thrust of the UAV; wherein, the physical characterization of the manifold adsorption index is the ratio of the centripetal negative pressure gradient to the equivalent dynamic pressure of the air density and the upper limit of the maximum wind resistance thrust in the fire; marking the continuous set of spatial grids with a manifold adsorption index greater than the preset critical adsorption threshold and a continuously increasing tangential wind speed vector as the manifold adsorption trap zone.

3. The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion according to claim 1, characterized in that: The specific steps for calibrating the extreme thermal convection divergence zone include: extracting the vertical heat flux directly above the center of the forest crown fire based on real-time environmental meteorological and combustion heat field distribution data; calculating the predicted vertical updraft velocity within the vortex eye directly above the fire based on the thermodynamic plume core region equation; calculating the predicted terminal free settling velocity of the mission payload at low air density at various height levels in the fire based on the mass, windward cross-sectional area, and air resistance coefficient of the mission payload carried by the core operation UAV; calculating the unit-time heat evaporation loss rate of the extinguishing payload as it falls through the spatial grid based on the thermal radiation law; and jointly marking the three-dimensional spatial set where the predicted vertical updraft velocity is greater than or equal to the predicted terminal free settling velocity of the payload, and the heat evaporation loss rate of the payload is greater than or equal to the preset maximum allowable loss rate threshold, as the extreme thermal convection divergence zone, and marking the bottom spatial layer that meets the dual failure conditions as the upper boundary of the divergence zone.

4. The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion according to claim 1, characterized in that: The specific steps to trigger the first asymmetric cluster penetration strategy include: obtaining the entry point of the core operation drone into the manifold adsorption trap zone and reading the wind direction in the absolute coordinate system in real time; deploying an asymmetric topological formation with the core operation drone as the origin of the coordinate system in the opposite direction of the wind, wherein the density of the escort drones on the windward side facing the tangential main wind flow of the vortex center is greater than the density on the leeward side away from the vortex center; based on the flow tube momentum theory and the law of conservation of mass, introducing the wake expansion coefficient dynamically coupled with the measured temperature of the fire site and the reference atmospheric temperature, calculating the wind speed deficit generated after the rotor of a single escort drone extracts air kinetic energy; superimposing the wind speed deficit generated by each escort drone in the asymmetric topological formation to generate an equivalent local low wind speed stable flow field three-dimensional space biased towards the windward side for pressure resistance, which is defined as the asymmetric three-dimensional aerodynamic shielding envelope; extracting the solid three-dimensional physical vehicle point cloud set of the core operation drone, performing real-time spatial topological inclusion judgment at high frequency, and establishing the core constraint condition that the three-dimensional physical boundary of the core operation drone must be completely enveloped within the safe isosurface boundary of the asymmetric three-dimensional aerodynamic shielding envelope.

5. The method for cooperative scheduling of low-altitude airspace UAV swarms based on multi-source data fusion according to claim 1, characterized in that: After establishing the core constraints: once the fire field gusts cause the aerodynamic shielding safety envelope boundary to contract inward, and the physical boundary of the core operation UAV is detected to have escaped the safe area, the windward side escort UAV is prioritized to increase the current pitch angle of attack, and the thrust coefficient is increased to increase the axial induction factor to extract air momentum; while increasing the pitch angle of attack, spatial topology compensation is performed simultaneously, and the downwind axial distance between the windward side escort UAV and the core operation UAV is shortened according to a preset ratio to obtain the equivalent gain of shielding wind speed, until it forcibly penetrates the outer edge of the fire line.

6. The low-altitude airspace UAV swarm cooperative scheduling method based on multi-source data fusion according to claim 1, characterized in that: In the second tangential slingshot escape and cluster formation self-healing strategy, the process of blocking conventional flight control and executing acceleration includes: when the manifold adsorption index of any escort UAV exceeds the limit bearing threshold, the conventional flight control system's headwind deceleration and reverse hovering commands are hard blocked; the tangential wind speed vector of the wind field at the current coordinates of the endangered UAV and the radial distance relative to the vortex center are obtained, and the minimum composite velocity to maintain non-inhalation is calculated by combining its own mass and thrust as the centrifugal escape critical value. The physical principle of this critical value is to use the centrifugal force generated by the tailwind acceleration to counteract the centripetal negative pressure suction of the fire field; the endangered UAV is controlled to apply the maximum positive thrust in the positive direction of the tangential wind speed vector to generate tangential smooth acceleration, and the absolute velocity of the fuselage and the tangential wind speed of the fire field are superimposed in the forward vector to obtain the composite motion velocity.

7. The low-altitude airspace UAV swarm cooperative scheduling method based on multi-source data fusion according to claim 1, characterized in that: The specific steps to complete the tangential slingshot escape are as follows: When the ratio of the square of the combined velocity to the radius of the vortex center satisfies the critical value condition for centrifugal escape, the thrust vector ejection angle is calculated based on the topological curvature of the three-dimensional thermodynamic flow field. The calculation logic of the thrust vector ejection angle is as follows: the maximum critical sideslip angle corresponding to the aerodynamic shape of the aircraft is extracted as the physical upper limit reference value; the ratio of the vortex core radius to the real-time centripetal radius of the current UAV position is extracted as the flow field spatial geometric topological structure parameter; this structural parameter is substituted into the trigonometric function nonlinear mapping relationship to dynamically output a continuous constraint degree interval that deviates outward; the UAV is controlled to output thrust at full power according to the thrust vector ejection angle, and the centrifugal force field of the fire tornado field is used to throw the UAV out of the boundary of the manifold adsorption trap area along the parabolic trajectory to complete self-protection.

8. The method for cooperative scheduling of low-altitude airspace UAV swarms based on multi-source data fusion according to claim 1, characterized in that: The specific steps for synchronously scheduling backup drones to replace the formation include: extracting the three-dimensional coordinates of the spatial topological gaps left by the escort drones that have broken away from the formation in the asymmetric formation; calculating the dynamic manifold replacement rate and the corresponding limit safety time window based on the destruction rate of high-temperature fluid in the fire field penetrating into the aerodynamic shielding boundary through the gap, combined with the physical width of the formation gap and the thermal expansion safety factor; within the limit safety time window, immediately scheduling a backup escort drone in a safe redundancy state from the leeward side of the weak pressure zone of the asymmetric formation, crossing the top or rear of the core drone at maximum maneuver speed to diagonally cut in, accurately filling the spatial topological gaps on the windward side, and maintaining the aerodynamic shielding volume from physical rupture.

9. The low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion according to claim 1, characterized in that: The specific steps of the second operational phase, namely the negative pressure suction and throwing, include: after the core operational UAV enters the extreme thermal convection divergence zone, the windbreak formation of the first operational phase is released, and the escort UAV switches to a wave-riding warning attitude that utilizes the rising thermal airflow for lift; the core operational UAV reads in real time the vertical rising airflow velocity and the central negative pressure value of the core area of ​​the three-dimensional vortex tube directly below; the predicted value of the free settling velocity at the end of the fire extinguishing load calculated when determining the area is retrieved and compared in real time, and the throwing lock is released only when the predicted value of the free settling velocity at the end is strictly greater than the real-time vertical rising velocity; the core operational UAV is controlled to aim at the geometric negative pressure center of the thermal convection divergence zone for delivery, and a time-delay fuse is set to detonate at the phase change section of the vortex tube airflow, using the absolute negative pressure funnel effect at the center of the fire tornado rotation to reversely draw the detonated fire extinguishing agent downward to the core deflagration zone.

10. A low-altitude airspace UAV swarm collaborative scheduling system based on multi-source data fusion, used to execute the low-altitude airspace UAV swarm collaborative scheduling method based on multi-source data fusion as described in any one of claims 1-9, characterized in that: include: The multi-source fusion and airspace topology reconstruction module is used to acquire a three-dimensional environmental topology map of the fire site and UAV dynamic parameters, fuse radar and multispectral data, perform spatial field modeling, and calibrate the three major constraint boundaries of the three-dimensional vortex core region, manifold adsorption trap region, and extreme thermal convection divergence region. The path cross-validation and intelligent diversion module is used to extract the initial trajectory of the UAV and perform overlap verification of trap zone nodes. It generates detour and obstacle avoidance corridors for routine tasks and triggers special operation strategy switching for core tasks. The asymmetric aerodynamic shielding dynamic generation module is used to schedule the poses of multiple machines in the first operation phase, calculate the wind speed deficit based on the momentum conservation and wake thermal expansion mechanism, and generate and compensate the low wind speed safety envelope volume of the eccentric structure in real time. The manifold adsorption monitoring and slingshot self-healing module is used to monitor the airframe's load-bearing limit at high frequency, hard block conventional flight control triggers tangential smooth acceleration, map the escape and jettison angle according to the topological curvature, and schedule redundant airframes to fill the gaps in the formation according to the penetration and damage rate. The flow field game and negative pressure throwing control module is used to take over the flight control in the second operation phase, execute the cluster disbanding and wave-riding gliding commands, and control the payload to perform precise release and delayed detonation towards the center of the vortex negative pressure funnel by comparing the settling velocity with the dynamics of the updraft.

Images